Prediction of the mechanism of miRNAs in laryngeal squamous cell carcinoma based on the miRNA-mRNA regulatory network

Sample source

LSCC miRNA and mRNA chip data were collected from the GSE124678 and GSE59102 retrieved from the National Center for Biotechnology Information (NCBI) GEO Datasets (http://www.ncbi.nlm.nih.gov/geo). Derived from the GPL16770 platform, GSE124678 includes five normal tissue samples and 32 tumor tissue samples. Derived from the GPL6480 platform, GSE59102 includes 13 normal tissue samples and 29 tumor tissue samples. The miRNA and mRNA microarray data of LSCC and normal laryngeal tissues were processed using platform annotation file (Edgar, Domrachev & Lash, 2002). We then used PERL5.30.2 (https://www.PERL.org/) to analyze and process the dataset; according to the annotation platform file of the expression profile, the probe name was converted to the corresponding gene name, and the empty probe was removed. Then, the miRNA and mRNA data were obtained by sequencing the data from the two datasets in order of the normal group and the tumor group.

Analysis of differentially expressed miRNAs and mRNAs

We used the LIMMA software package from R software (https://rstudio.com/, ver. 3.6.2). Objective to screen miRNAs and mRNAs differentially expressed in LSCC (Colaprico et al., 2016) The screening criteria we set are: —log2(fold change)—>1 and FDR (False Discovery Rate)<0.05. Then, the “pheatmap” package and “ggplot2” package in R (3.6.2) were used to produce the volcano map. The differentially expressed miRNAs and mRNAs were screened.

Prediction of miRNA targets

The effects of miRNAs are mediated by their complete or incomplete interactions with target genes, affect the expression of target genes. Hence, miRNAs target genes must be predicted. In this study, 99 differentially expressed miRNAs target genes were predicted by FunRich (3.1.3).

Construction of the miRNA-mRNA regulatory network diagram

First, we used PERL software to intersect the target genes of miRNAs with the differentially expressed mRNAs in LSCC to obtain the common mRNAs and corresponding miRNAs. Previously, a negative regulatory relationship between miRNAs and their target genes was identified (Bartel, 2009). Therefore, based on the negative regulatory relationship between miRNAs and mRNAs, miRNA-mRNA pairs with significant differences in log2 (fold change) values were selected, and Cytoscape (3.7.2) software was used to visualize the regulatory network between the two to obtain the miRNA-mRNA regulatory network (Lin & Chen, 2018; Zhao et al., 2020).

GO and KEGG enrichment analyses

We performed GO and KEGG enrichment analyses of differentially expressed mRNAs to further clarify the biological functions of target genes in the regulatory network (Kanehisa & Goto, 2000; Khatri, Sirota & Butte, 2012). R software in the Bioconductor plug-in (http://www.bioconductor.org) was used to perform the GO and KEGG enrichment analyses with the criteria of p < 0.05 and q < 0.05. The results are given in the form of bar chart.

Cell culture and cell transfection

We used Human LSCC cell line LSC-1 (Bluefbio, China) and human laryngeal epithelial cells (HLEC; Lifeline, America). Dulbecco’s Modified Eagle’s medium (DMEM) was used as the basic medium and 10% foetal bovine serum (FBS) and 100 g/ml penicillin/streptomycin were added to the medium, the culture conditions were 37 °C and 5% CO₂. In this study, the cells were all used at passages 2–4 after recovery.

A total of 5 nmol miR-140-3p miRNA mimics (miR-140-3p) or miRNA mimic negative control (miR-NC) (GENECREATE, Wuhan, China) was dissolved in 0.9%NaCl as a storage solution of 20 µM. LSC-1 cells (3*105 cells per well) were cultured in 6-well plates overnight. When cell density reached 60–70%, the cell medium was replaced with a complete medium without antibiotics. 1.5 µL of miR-140-3p or miR-NC storage solution was diluted with 50 µL serum-free medium, and 1 µL Lipofectamine²⁰⁰⁰ (Beyotime, Shanghai, China) was added to the other 50 µL serum-free medium. Mix well and let stand at 37 °C for 5 min. Mix the two tubes together and let stand at 37 °C for 20 min. Then add the mixture into each cell well, gently shake the culture plate to mix the complex with the cells, add 500 µL complete medium, and place in the cell incubator for further culture for 48 h, then collect the cells for subsequent experiments. All cells were cultured in complete medium for at least 24 h before transfection and rinsed with phosphate-buffered saline (PBS, pH 7.4) before transient transfection. The generated cell clones were tested for stable miR-140-3p expression.

RNA extraction and QRT- PCR

According to the manufacturer’s instructions, total RNA was extracted from the cell using Trizol reagent (TAKARA, Japan). Total RNA was reverse transcribed into cDNA using the PrimeScript RT Reagent Kit (Takara, Japan). Reverse transcription: 25 °C10 min, 50 °C for 30 min and 85 °C for 5 min. Expression was detected using the fluorescence quantitative PCR kit and the following conditions: there are 40 cycles at 95 °C for 5 min, 95 °C for 10 s, 60 °C for 30 s. The solubility curve temperature was set to 60−95 °C, and three replicate wells were set for each specimen. The expression of miR-140 was normalized to a small nuclear RNA (U6) as an internal reference. The results were calculated using the 2^−ΔΔCt method. The primer sequences used for quantitative real-time PCR analyses of miR-140-3p were as follows: forward, 5′-ACACTCCAGCTGGGTACCACAGGGTAGAA-3′ and reverse, 5′- CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCCGTGGTT-3′. The primer sequences for small nuclear RNA (u6) are as follows: forward, 5′- CTCGCTTCGGCAGCACA -3′ and reverse 5′- AACGCTTCACGAATTTGCGT -3′. The relative expression of miRNA was detected by 2^−ΔΔCt.

CCK8 assay

We performed the CCK8 assay to measure the proliferation of LSC-1 cells. Cells were cultured in 96-well culture plate, and the inoculation density was 3*10⁵/well (Jiang et al., 2019). For cell transfection, cells were cultured overnight. After 48 h of transfection, 10 mL of CCK8 solution was added to each well, and the cells were incubated at 37 °C for another 60 min. The absorbance of the solution was measured at 490nm by Smart Microplate Reader (SMR) 16.1.

Flow cytometry analysis

LSC-1 cells were transfected for 48 h. After transfection with miR-140-3p mimics or control miRNA, LSC-1 cells were washed with cold PBS buffer. Then, LSC-1 cells were detached with trypsin and washed twice with cold PBS. According to the manufacturer’s instructions, cells were subsequently double stained with FITC-labelled Annexin V and propidium iodide (PI) using the FITC Annexin V Apoptosis Detection Kit (BD Biosciences, USA). Flow cytometer was performed to determine the percentage of apoptotic cells (FACS Calibur; BD Biosciences, USA). The data were analyzed by FLOWJ software.

Statistical analysis

GraphPad Prism5.0 and statistical product and service solutions (SPSS) 22.0 software were used for statistical analyses. Use t-test to enumeration data, and P < 0.05 indicated the significance level.

Differentially expressed miRNAs and mRNAs

After annotating the GSE124678 dataset, 1205 pieces of miRNA information were obtained from human laryngeal carcinoma tissues, including 5 tissues in the normal group and 32 tissues in the tumour group. 99 differentially expressed miRNAs were screened by PERL language (Table S1 for details). A volcanic map of differentially expressed miRNAs is then drawn using R software (Fig. 1). Similarly, we used the GSE59102 dataset from the LSCC mRNA chip downloaded from the GEO database to create a volcanic map of the differentially expressed mRNAs for human LSCC (Fig. 2). Thirteen tissues in the normal group and 29 tissues in the tumour group were analysed. A total of 2758 differentially expressed mRNA, were obtained, including 1312 up-regulated and 1446 down-regulated (Table S2 for details).

Prediction of miRNA targets

FunRich (3.1.3), a gene function analysis tool, was used to predict the target genes of the 99 differentially expressed miRNAs; 1386 target genes were obtained in the background, and the number of miRNAs matched with them was 32.

Construction of the miRNA-mRNA network diagram

The differentially expressed mRNAs and predicted miRNA target genes in laryngeal carcinoma were processed using PERL. The intersection of miRNA target gene and mRNAs, that is, the targeted regulation relationship between miRNAs and mRNAs, was obtained. Then, Cytoscape software was used to map the miRNA-mRNA regulatory network (Fig. 3). This miRNA-mRNA regulatory network contains 10 miRNAs and 96 mRNAs (Table S3 for details). The more genes that are connected with miRNAs in this regulatory network indicate that this miRNA is very important in the occurrence and development of LSCC. As this regulatory network clearly shows, miR-140-3p regulates more mRNAs. It is suggested that miR-140-3p plays an important role in the occurrence and development of LSCC.

Functional enrichment analysis

We conducted GO and KEGG enrichment analyses of the 2758 differentially expressed mRNAs using p < 0.05 and q < 0.05 as screening conditions. The GO analysis includes three categories: molecular function (MF), cellular component (CC), and biological process (BP). These differentially expressed mRNAs promoted the occurrence and development of LSCC by participating in various BPs, CCs and MFs (Fig. 4). Similarly, the results of KEGG enrichment analysis were also included Cell cycle (hsa04110), p53 signaling pathway (hsa04115), IL-17 signaling pathway (hsa04657), chemical carcinogenicity (hsa05204), etc (Fig. 5, Table 1).

The expression of miR-140-3p in LSC-1 cells is low

In this study, the expression of miR-140-3p in LSC-1 cells and HLECs was detected using QRT- PCR. We examined the expression of miR-140-3p in HLECs and LSC-1 cells, and observed lower miR-140-3p expression in LSC-1 cells than in HLECs. Obviously, these experimental results suggest that the decrease in miR-140-3p expression in LSC-1 cells. The results were consistent with bioinformatics analysis (Fig. 6).

Upregulation of miR-140-3p expression inhibits LSC-1 cells proliferation in vitro

The CCK8 assay was performed to further assess the biological role of miR-140-3p in LSC-1 cells. LSC-1 cells were transfected with miR-NC or miR-140-3p mimics. Cell proliferation was detected by CCK8 method. QRT- PCR results showed that the transfection of miR-140-3p mimics led to overexpression of miR-140-3p in LSC-1 cells compared with the control group (Fig. 7). Similarly, the results of the CCK8 assay showed significantly impaired growth of LSC-1 cells transfected with miR-140-3p mimics compared with LSC-1 cells transfected with miR-NC. Based on these results, the high expression of miR-140-3p in LSC-1 cells inhibited their proliferation (Fig. 8).

The upregulation of miR-140-3p promoted the apoptosis of LSC-1 cells

Flow cytometry was used to further clarify the effect of miR-140-3p upregulation on the apoptosis of LSC-1 cells. Compared with miR-NC, miR-140-3p mimics significantly promoted the apoptosis of LSCC cells. Thus, the upregulation of miR-140-3p promoted the apoptosis of LSC-1 cells (Fig. 9).